Sparse matrix import and Parallelization

With this release of HiCBricks,

• I have implemented import of sparse matrices (upper-triangle non-zero values).
• The package now supports parallel retrieval operations. Not to be confused with traditional parallel operations, HiCBricks allows users to parallelize operations across all chromosome pairs. Within matrix parallel operations are not supported at this time.

HiCBricks and S4 classes (Bioconductor 3.10)

With this release of HiCBricks, the objective was to

• Decrease I/O time overheads associated with storing large square matrices.
• Allow the storage of resolutions.
• To streamline the entire workflow by removing parameters.
• To allow exporting of data stored in Brick format.

I re-imagined the entire format in terms of S4 classes, which provide incredible value to Bioconductor packages. Hi-C matrices are no more stored together, but separately. HiCBricks, now creates only upper triangle chromosome pair matrices.

They are now tracked inside R sessions using an S4 class. In the folder location they are tracked via a JSON file.

Parameters with little functionality, example, exec have been removed. Others are also on the list.

Export to sparse format has also been added.

Releasing the first github release

This is a stable release of HiCBricks, from here on out the accessors will remain stable. No guarantees are made for the format, which may be expanded in a future v2 release.

• Accessors are stable
• Vizart plotting functions are stable
• Vignette is reasonably well-written